Open sonnystalin opened 2 months ago
It looks like the shapes of the tensor x and img do not match, and you can print their shapes for further debugging.
I used my own data set for testing. The picture is 2-dimensional, but in the end there is always one more dimension and the value is 0.
What parameters need to be changed to run your own data set using Windows system?
Excuse me, have you ever processed the data set? Or use it directly? If you have done dimensionality reduction or other processing, could you please publish it? Thank you for your help sincerely! @takimailto
Sorry, we have no further preprocessing. You can give me the shape of your data loader and I can use it to test.
Help, how to solve this question? unet.py", line 492, in forward x = th.cat([x, img], dim=1) RuntimeError: Tensors must have same number of dimensions: got 4 and 5